search

hytseng0509 / CrossDomainFewShot

Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation (ICLR 2020 spotlight)
331 stars 62 forks source link

About the pre-trained backbone #2

Closed yxgnahz closed 4 years ago

yxgnahz commented 4 years ago

Hello, Thanks for your work! I have one question for the pre-trained Resnet-10 backbone: did you use different hyper parameters to train this backbone or we can just get the same by running the code "train_baseline.py" where I see you use default Adam optimizer training 400 epochs?

hytseng0509 commented 4 years ago

Hi,

We use the default hyper-parameters to train the pre-trained feature encoder (i.e., Adam, 400 epochs).